1 / 24

Overview of Kernel Methods

Overview of Kernel Methods. Prof. Bennett Math Model of Learning and Discovery 2/27/05 Based on Chapter 2 of Shawe-Taylor and Cristianini. Outline. Overview Ridge Regression Kernel Ridge Regression Other Kernels Summary . Kernel Method. Two parts (+):

demetria
Download Presentation

Overview of Kernel Methods

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Overview of Kernel Methods Prof. Bennett Math Model of Learning and Discovery 2/27/05 Based on Chapter 2 of Shawe-Taylor and Cristianini

  2. Outline • Overview • Ridge Regression • Kernel Ridge Regression • Other Kernels • Summary

  3. Kernel Method Two parts (+): • Mapping into embedding or feature space defined by kernel. • Learning algorithm for discovering linear patterns in that space. • (Leaning algorithm must be regularized.) • (Learning algorithm must work in dual space.) Illustrate using linear regression.

  4. Linear Regression Given training data: • Construct linear function: • Creates pattern function:

  5. 1-d Regression

  6. Predict Drug Absorption • Human intestinal cell line CACO-2 • Predict permeability • 718 descriptors generated • Electronic TAE • Shape/Property (PEST) • Traditional • 27 molecules with tested permeability

  7. Least Squares Approximation • Want • Define error • Minimize loss

  8. Optimal Solution • Want: • Mathematical Model: • Optimality Condition: • Solution satisfies: Solving nn equation is 0(n3)

  9. Solution • Assume exists, then • Alternative Representation: Is this a good assumption?

  10. Ridge Regression • Inverse typically does not exist. • Use least norm solution for fixed • Regularized problem • Optimality Condition: Requires 0(n3) operations

  11. Ridge Regression (cont) • Inverse always exists for any • Alternative representation: Solving ll equation is 0(l3)

  12. Dual Ridge Regression • To predict new point: • Note need only compute G, the Gram Matrix Ridge Regression requires only inner products between data points

  13. Efficiency • To compute w in primal ridge regression is 0(n3)  in primal ridge regression is 0(l3) • To predict new point x primal 0(n) dual 0(nl) Dual is better if n>>l

  14. Notes on Ridge Regression • “Regularization” is key to addressstability and regularization. • Regularization lets method work when n>>l. • Dual more efficient when n>>l. • Dual only requires inner products of data.

  15. Linear Regression in Feature Space Key Idea: Map data to higher dimensional space (feature space) and perform linear regression in embedded space. Embedding Map:

  16. Nonlinear Regression in Feature Space In primal representation:

  17. Nonlinear Regression in Feature Space In dual representation:

  18. Kernel Function • A kernel is a function K such that • There are many possible kernels. Simplest is linear kernel.

  19. Derivation of Kernel

  20. Ridge Regression in Feature Space • To predict new point: • To compute the Gram Matrix Use kernel to compute inner product

  21. Popular Kernels based on vectors By Hilbert-Schmidt Kernels (Courant and Hilbert 1953) for certain  and K, e.g.

  22. Kernels Intuition • Kernels encode the notion of similarity to be used for a specific applications. • Document use cosine of “bags of text”. • Gene sequences can used edit distance. • Similarity defines distance: • Trick is to get right encoding for domain.

  23. Important Points • Kernel method = linear method + embedding in feature space. • Kernel functions used to do embedding efficiently. • Feature space is higher dimensional space so must regularize. • Choose kernel appropriate to domain.

  24. Kernel Ridge Regression • Simple to derive kernel method • Works great in practice with some finessing. • Next time: Practical issues. More standard dual derivation.

More Related